Team Ai
Datasetpublic

openadmet/cyp-challenge-train-test

CYP Challenge Train/Test Dataset A high-quality experimental dataset for predicting inhibition of the major drug-metabolizing Cytochrome P450 enzymes (CYP1A2, CYP2C9, CYP2D6, CYP3A4), released as part of the OpenADMET CYP Inhibition Blind Challenge. Blog post: Announcing OpenADMET’s CYP inhibition blind challenge Challenge Space: OpenADMET CYP Inhibition Blind Challenge Challenge period: August 17, 2026 - November 3, 2026 Produced by: OpenADMET CHANGELOG Updated… See the full description on the dataset page: https://huggingface.co/datasets/openadmet/cyp-challenge-train-test.

sourceHugging Faceapache-2.0updated 17d agoView on Hugging Face
12likes3.2kdownloads
Dataset Card

CYP Challenge Train/Test Dataset

A high-quality experimental dataset for predicting inhibition of the major drug-metabolizing Cytochrome P450 enzymes (CYP1A2, CYP2C9, CYP2D6, CYP3A4), released as part of the OpenADMET CYP Inhibition Blind Challenge.

Blog post: Announcing OpenADMET’s CYP inhibition blind challenge

Challenge Space: OpenADMET CYP Inhibition Blind Challenge

Challenge period: August 17, 2026 - November 3, 2026

Produced by: OpenADMET

CHANGELOG

  • —Updated 2026-09-23 Added the structure track's blinded test set (SMILES for the co-folding/structure-prediction task; ground-truth CYP3A4 complex structures used for scoring are not disclosed yet).
  • —Updated 2026-08-06 Finalised ground truth: refreshed direct-inhibition and TDI training/test data with the finalised assay results, and added a new Emax (maximal effect) training set.
  • —Updated 2026-08-03 Initial release: direct inhibition training/blinded test data, time-dependent inhibition (TDI) training data, and single-concentration screening training data.

Dataset contents

ConfigSplitFileDescription
defaulttraincyp-challenge-TRAIN_inhibition.csvPrimary direct-inhibition training set — 4,905 compounds with pIC50 (plus 95% CI and std) for CYP1A2, CYP2C9, CYP2D6, CYP3A4
defaulttestcyp-challenge-TEST-BLINDED.csv750-compound blinded test set (SMILES only; labels withheld for the challenge)
tditraincyp-challenge-TRAIN_TDI.csvTime-dependent inhibition (TDI) training set — 6,145 compounds with is_TDI classification labels for CYP2D6/CYP3A4, TDI-condition pIC50s, and paired direct-inhibition pIC50s for comparison
single_concentrationtraincyp-challenge-single-concentration-TRAIN.csvSingle-concentration screening data — 17,504 measurements across 4,376 compounds x 4 enzymes (log2 fold-change format)
emaxtraincyp-challenge-TRAIN_Emax.csvEmax (maximal effect) training set — 6,146 compounds with is_TDI classification labels for all four CYPs, plus TDI-condition and direct-inhibition Emax values (with 95% CI)
structuretestcyp-challenge-TEST-BLINDED_structures.csv20-compound blinded test set for the structure-prediction track — Molecule_Name, SMILES only; the ground-truth CYP3A4-ligand complex structures used for scoring are withheld for the challenge

Loading with Hugging Face datasets

python
from datasets import load_dataset

# Default config (primary direct-inhibition assay)
ds = load_dataset("openadmet/cyp-challenge-train-test")
train = ds["train"]
test  = ds["test"]

# TDI (time-dependent inhibition) config
ds_tdi = load_dataset("openadmet/cyp-challenge-train-test", "tdi")
train_tdi = ds_tdi["train"]

# Single-concentration config
ds_single = load_dataset("openadmet/cyp-challenge-train-test", "single_concentration")
train_single = ds_single["train"]

# Emax config
ds_emax = load_dataset("openadmet/cyp-challenge-train-test", "emax")
train_emax = ds_emax["train"]

# Structure config
ds_structure = load_dataset("openadmet/cyp-challenge-train-test", "structure")
test_structure = ds_structure["test"]

Loading directly with pandas

python
import pandas as pd

train        = pd.read_csv("hf://datasets/openadmet/cyp-challenge-train-test/cyp-challenge-TRAIN_inhibition.csv")
test         = pd.read_csv("hf://datasets/openadmet/cyp-challenge-train-test/cyp-challenge-TEST-BLINDED.csv")
train_tdi    = pd.read_csv("hf://datasets/openadmet/cyp-challenge-train-test/cyp-challenge-TRAIN_TDI.csv")
train_single = pd.read_csv("hf://datasets/openadmet/cyp-challenge-train-test/cyp-challenge-single-concentration-TRAIN.csv")
train_emax   = pd.read_csv("hf://datasets/openadmet/cyp-challenge-train-test/cyp-challenge-TRAIN_Emax.csv")
test_structure = pd.read_csv("hf://datasets/openadmet/cyp-challenge-train-test/cyp-challenge-TEST-BLINDED_structures.csv")